This is a petition for review of a Federal Energy Regulatory Commission ruling on a request by Transcontinental Gas Pipeline (TGPL) for a rate increase under Section 4 of the Natural Gas Act, 15 U.S.C. Sec. 717c (1988). Petitioners (hereinafter NCUC), the regulatory agencies of North Carolina and New York, and PECO Energy Company, a utility providing gas and electric service in eastern Pennsylvania and northern Maryland, challenge an order of the Federal Energy Regulatory Commission allowing TGPL to earn an alleged return of 24% on its common equity. Specifically, petitioners challenge the Commission's use of a hypothetical capital structure to reach its result. They also challenge the Commission's decision to select a rate of return at the high end of the zone of reasonableness for a natural gas pipeline's common equity. In a separate challenge, petitioner TGPL contends the Commission erred in employing a proxy group of pipeline parent companies to derive the hypothetical structure. Because we find FERC's decision arbitrary and capricious on these issues, we reverse and remand.

The Federal Energy Regulatory Commission (FERC) establishes the rates of a regulated pipeline by determining the pipeline's total revenue requirements. Such requirements are composed of the pipeline's rate base, its operating costs, and a rate of return sufficient to ensure that pipeline investors are fairly compensated. The rate of return component is calculated based on the weighted average of the costs of the three elements comprising the capital structure: debt, preferred stock, and common equity. To arrive at the proper rate of return, the Commission first determines the appropriate capital structure for the pipeline. The Commission then sets a rate of return for each element of the capital structure. Since the returns attributable to the three components comprising the overall return differ markedly, the capital structure used by the regulated entity to finance its investment can greatly affect the overall return allowed.

The area of contention in this case is the return allowed on the common equity element. To formulate a return on this component, the Commission develops a zone of reasonableness based on the range of returns generally experienced in the industry. The Commission then adjusts the return of the particular pipeline at issue within the zone to reflect the specific investment risks of that pipeline as compared to similar investments.

The proceedings below were initiated on March 2, 1992, when TGPL, a natural gas pipeline subject to FERC's rate regulation, filed new tariff sheets with the Commission pursuant to Section 4(e) of the Natural Gas Act, 15 U.S.C. Sec. 717c(e). TGPL desired to increase its rates for natural gas service by $234 million. The largest single component of the rate increase was TGPL's request for a higher rate of return on its investment, specifically an after tax return of 36.4% on its common equity.

Because the Commission had traditionally allowed returns on common equity in the 11%-15% range, TGPL's request produced much protest from TGPL's customers and affected state commissions. As a result, the Commission issued an order not only suspending the rate increase until September 1, 1992, but also establishing an expedited hearing to consider the appropriate rate of return on common equity. Transcontinental Gas Pipe Line Corp., 59 FERC p 61,034 at 61,102 (1992).

On July 10, 1992, the administrative law judge (ALJ) issued an order rejecting TGPL's proposal for a 36.4% rate of return on equity. The ALJ concluded the proposal relied on unacceptable methodologies and lacked adequate evidentiary support. Transcontinental Gas Pipe Line Corp., 60 FERC p 63001 at 65,044 (1992). Specifically, the ALJ concluded TGPL had erred in basing its rate of return proposal on the capital structure of its parent, Transco Energy Company (TEC), rather than on TGPL's own capital structure. 60 FERC at 65,028-34. The ALJ also rejected an alternative rate of return proposed by petitioner Public Service Commission of the State of New York because it was based on TEC's capital structure and used an inappropriate comparison group to set the rate of return: local natural gas distribution companies. Id. at 65,047-48.

The ALJ ultimately adopted a rate of return proposal advanced by financial analyst George Shriver on behalf of the Commission staff (hereinafter the Shriver analysis). Consistent with the ALJ's determination that the pipeline's capital structure should be used, the Shriver analysis was based on TGPL's capital structure. Id. at 65,048. To determine the zone of reasonableness for TGPL's rate of return, Shriver used a primary comparison group of seven publicly owned pipeline parent corporations obtaining over 50% of their revenues from their natural gas pipeline subsidiaries, as well as a secondary comparison group of twenty natural gas pipeline companies that do not have publicly traded stock. Applying a Discounted Cash Flow (DCF) analysis to the secondary comparison group, Shriver arrived at a zone of reasonableness for a rate of return on equity ranging from 9.49% to 13.96%. Shriver further concluded that because TGPL was among the riskiest investments in the industry, TGPL's rate of return on equity should be set at 13.96%, the upper end of the zone of reasonableness. The ALJ found the Shriver analysis the best supported and methodologically soundest. Id. at 65,048-51.

TGPL and NCUC appealed the ALJ's order to the Commission. On September 17, 1992, the Commission issued an order adopting the conclusions of the ALJ, except for certain aspects of the rate of return calculation. Transcontinental Gas Pipeline Corp., 60 FERC p 61,246 at 61,820 (1992). The Commission reversed the ALJ's decision to use TGPL's actual capital structure. The Commission acknowledged under normal circumstances, its policy would be to impute TEC's capital structure to TGPL because TEC is the financing source for TGPL's operations, as TGPL raises none of its own capital directly. Id. at 61,823. However, the Commission found that TEC's common equity ratio of 16.27% was atypically low. The Commission cryptically declared, as a result, that TEC would require an anomalously high rate of return in relation to rates of return approved for comparable pipelines. Id. Consequently, the Commission determined TGPL's rate of return on common equity should be derived using a hypothetical capital structure. Id.

Because none of the witnesses in the evidentiary hearings testified in support of utilizing a hypothetical capital structure, the Commission relied on the comparison groups discussed in the Shriver analysis. The Commission found the primary comparison group of seven publicly owned companies with pipeline subsidiaries was the most appropriate comparison group to establish a hypothetical capital structure because these publicly traded companies, like TEC, derived more than 50% of their revenue from pipeline corporations. Id. at 61,824. The average common equity ratio of these companies was 38.79%. The Commission excluded from this proxy group the secondary comparison group of 20 privately held pipeline companies because their stock was not publicly traded. Id.

Based on this hypothetical capital structure, the Commission proceeded to calculate TGPL's rate of return on equity. The Commission first rejected the ALJ's use of the secondary support group in establishing a zone of reasonableness. The Commission asserted that market data is the linchpin of DCF analysis. Consequently, because the secondary comparison group was comprised of privately owned pipelines, whose stock was not publicly traded, the Commission rejected it. Id. at 61,826. Relying on the primary comparison group of publicly held, unregulated parent companies, the Commission determined the zone of reasonableness for TGPL's rate of return on equity was between 9.2% and 14.45%. Id.

The Commission further found that a return on common equity at the top of the reasonableness zone was proper. The Commission concluded that TGPL's business and financial risks were higher than those of all other pipelines except for one, and TGPL's credit ratings were below investment grade. To account for these factors, the Commission set the return on equity at 14.45%, the high end of the curve. Id.

On July 8, 1993, the Commission issued an order denying the parties' petitions for rehearing. Transcontinental Gas Pipeline Corp., 64 FERC p 61,039 (1993). The Commission reaffirmed its earlier findings regarding the appropriate capital structure and rate of return on common equity for TGPL. Id. at 61,343, 61,348. The Commission elaborated on its reasoning for relying on the parent companies to base a hypothetical capital structure. The Commission stated the capital structure should be based on an actual capital structure developed in response to the need to attract capital in the financial markets. The secondary group of subsidiary pipelines, unlike the primary group, was removed from the realities of the financial markets where the capital to finance TGPL's operations must be raised. Id. at 61,345-46. Petitioners sought review of that order.

We review this order under the deferential standard mandated by section 706 of the Administrative Procedure Act, providing that a court must uphold a final agency action unless that action is "arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law." 5 U.S.C. Sec. 706(2)(A) (1988). The court must examine the Commission's reasoning to determine whether it considered the relevant factors and drew a rational connection between the facts found and the choice made. Associated Gas Distributors v. F.E.R.C., 824 F.2d 981, 1016 (D.C.Cir.1987). In addition, pursuant to Section 19(b) of the Natural Gas Act, the Commission's choice of a capital structure for TGPL must be based on substantial evidence. 15 U.S.C. Sec. 717r(b).

NCUC first contends the Commission's decision to adopt a hypothetical capital structure to determine TGPL's rate of return was arbitrary and capricious. The Commission purportedly failed to specify what rate of return it was assuming when it stated TGPL's rate would be anomalously high. Nor did the Commission indicate what "normal" rate it was using as a reference. Consequently, the Commission provided no explanation as to why the 14.45% rate it ultimately allowed was in fact appropriate and thus not anomalously high. More importantly, the Commission never explained a logical connection between avoiding a high return rate and the use of a hypothetical capital structure. According to NCUC, the Commission resorted to a fictional construct to mask a high return as something allegedly more reasonable.

FERC responds that its decision is well supported by the data demonstrating TGPL has an extremely low common equity ratio. For instance, the record demonstrates TEC's 16.27% common equity ratio is unusually low compared to the equity ratios of the primary comparison group, and thus requires an anomalously high rate of return to compensate for the increased riskiness of a correspondingly high debt ratio. 64 FERC at 61,346. In addition, pipeline equity ratios are ordinarily in the range of 50%. 64 FERC at 61,347 n. 24. In fact, the equity ratios contained in recent settlements involving TGPL itself were in the range of 30%. Finally, TEC's unusually low equity ratio is the result of recent debt financing for nonrecurring business losses and is thus not representative of TGPL's experience of risk. Id. at 61,346-47.

While these assertions may be true, we find FERC's justification inadequate. FERC never explained in its orders why the mere existence of an allegedly anomalously high rate of return must be avoided. The only explanation was supplied by TGPL in its brief as intervenor on behalf of FERC.1 TGPL stated the Commission did not want to set a precedent of granting an abnormally high rate of return on equity to TGPL, and thus force the Commission to distinguish this case when setting pipeline rates in future cases. By using a hypothetical capital structure, the Commission was able to use a lower rate of return on equity without lowering the total return dollars in TGPL's rates. At oral argument, FERC's counsel also adopted this explanation. However, this court cannot accept appellate counsel's post hoc rationalization of an agency decision. The Commission's decision "must be upheld, if at all, on the basis articulated by the agency itself." Motor Vehicle Mfrs. Ass'n v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 50, 103 S. Ct. 2856, 2870, 77 L. Ed. 2d 443 (1983) (citations omitted).

In any event, this rationale does not explain, for example, how a high rate of return acts as a precedent for future cases given the nature of the ratemaking process. According to FERC, the Commission was in the unenviable position of authorizing an anomalously high rate of return in this case only because TEC's common equity ratio was so low. Logically, FERC would only be in a similar position if it encountered a company with a similar common equity ratio. Yet by conceding TEC's common equity ratio was abnormally low compared to other companies, FERC also conceded the atypicality of this situation. If other companies suffer common equity ratios similar to TEC's, FERC has not explained the negative consequence of similarly allowing them high rates of return in comparison with other companies possessing higher common equity ratios and thus lower rates of return.

The sketchy justification for the rationale leaves unclear how the use of a hypothetical capital structure reduces the actual rate of return. NCUC argues the use of a hypothetical capital structure results in an actual return of 24.14% (38.35% pre-tax) on the common equity ratio that finances Transco's rate base. NCUC contends such a return is excessive because FERC had never previously allowed a major pipeline a similarly high return on common equity. Such a return, NCUC argues, violates F.P.C. v. Hope Natural Gas Co., 320 U.S. 591, 602, 64 S. Ct. 281, 287, 88 L. Ed. 333 (1944), holding the result reached must be just and reasonable. FERC responds that this argument is fallacious because it is based on the use of TEC's actual capital structure which produced anomalous results in the first instance. Rather, the rate of return is 14.45% when properly based on the hypothetical structure itself.

Without addressing whether a 24% actual rate of return is excessive, we find FERC's response problematic. FERC has not adequately explained how its reliance on the hypothetical capital structure to reach a 14.45% return is anything more than a device to mask an otherwise anomalous return as something more appealing. After all, according to FERC, individuals wishing to invest in TGPL desire a high rate of return to compensate for the perceived riskiness of investing in a pipeline operating under a higher debt load. Consequently, investors would necessarily focus on TEC's actual capital structure rather than FERC's fictitious structure when choosing to invest. FERC does not attempt to justify why its hypothetical is more than an elaboration of form over substance.

Before this court FERC relies on Communications Satellite Corp. v. FCC, 611 F.2d 883 (D.C.Cir.1977), as having approved the use of a hypothetical capital structure. Indeed, this is literally true. There, we approved the use of such a structure where a commission imputed hypothetical debt to reduce the allowable cost of capital to protect consumers since debt carries a lower cost than equity. That we have previously approved use of hypothetical capital structures under one circumstance does not by itself justify the use by a different commission of a different hypothetical for a different purpose. Here, FERC has imputed hypothetical equity to increase the cost of capital to protect investors. We hasten to make clear that we are not holding that FERC cannot do this. Indeed, FERC's prior decision in Kentucky West Virginia Co., 2 FERC p 61,139, at 61,325 (1978), may be a prior example of such use of a hypothetical capital structure. However, we do hold that the naked citation of prior authority for the use of a hypothetical under one circumstance does not automatically justify such in another. As we have held before, FERC has not provided us with adequate reasoning for review simply by citing prior authority "without explaining why that precedent was on point." Louisiana Interstate Gas Corp. v. FERC, 962 F.2d 37, 44 (D.C.Cir.1992) (emphasis in original). See also Maine Public Service Co. v. FERC, 964 F.2d 5, 9 (D.C.Cir.1992) (holding that FERC's use of a particular percentage in a ratemaking calculation was not adequately justified by citation of a prior use of the same percentage without further reasoning or explanation). Thus, if FERC wishes to proceed along these same lines on remand, it must provide further explanation of why it employed a hypothetical capital structure, and particularly why it employed the one that it used.

NCUC also contends, assuming the validity of the hypothetical capital structure, the Commission erred in allowing TGPL a rate of return at the top of the zone of reasonableness. The Commission concluded TGPL faced greater than average business and financial risks, and thus adopted a 14.45% return. NCUC asserts the Commission violated its own rule of engaging in risk determinations in a forward looking manner. The Commission failed to adjust for the reduction in business risks resulting from TGPL's switch in rate design on September 1, 1992, the date its increased rates in the proceeding below became effective. Prior to September 1992, TGPL's rates were designed on a modified fixed variable (MFV) basis, making the pipeline's recovery of equity return dependent on volumes of gas actually transported. However, effective September 1992, the company's rates were designed on a straight fixed variable (SFV) basis, thus assuring that the pipeline would recover its equity return through fixed charges independent of the volumes transported.

Consequently, NCUC maintains the Commission erred in relying on the alleged obsolete data of the Shriver study. Because the study was based on data for the 12-month period ending January 1992, the study necessarily compared TGPL's performance under MFV rates with the performance of other MFV pipelines. However, no evidence existed showing TGPL was currently unprofitable or faced any serious financial problems.

FERC responds that the financial markets were well aware of the Commission's decision to move to SFV rate design during the period scrutinized in the DCF analysis. The Commission's notice of proposed rulemaking proposing SFV rate design first issued on July 31, 1991. Also, the Commission required SFV rate design in Order No. 636, which issued on April 8, 1992. Thus, the markets knew that the SFV rate design would be imposed during the relevant study period. In any event, the Shriver analysis was based on the most recent data available at the time it was compiled. It incorporated all investor expectations regarding risk because it was based on the prices paid for the stock of TEC and other comparable enterprises. NCUC attempts to enumerate positive developments, such as the future strength of TGPL's markets. Both the Commission and the ALJ, however, concluded that mere optimism about the future is not sufficiently concrete to form the basis for establishing the rate of return. 60 FERC at 61,827.

We find the Commission's explanation does not adequately account for the shift to SFV. First, although the data underlying the DCF analysis reflected at least some anticipation of the change to SFV, it did not reflect its certain adoption. Because Order No. 636 was not issued until after the study's completion, for instance, it may not have been reflected in the financial markets' valuation of the sample group. Second, and more importantly, no account was taken of the impact that the switch to SFV would have on TGPL's relative riskiness--the factor truly at issue in choosing a place in the zone of reasonableness.

In addition to challenging the Commission's reliance on TGPL's business risk to place TGPL at the top of the zone of reasonableness, NCUC challenges the Commission's reliance on TGPL's financial risk, arguing the financial risk was mathematically eliminated by the adoption of a hypothetical capital structure. With an average capital structure, the requisite financial risk adjustment within the range of possible returns is zero. The hypothetical thicker equity ratio represents compensation for financial risk. Increasing the return on equity to the high end of the zone of reasonableness thus amounts to double dipping. FERC, however, maintains the imputation of such a structure places TGPL at a significant disadvantage because the hypothetical capital structure is an ideal capital structure. TGPL, however, must face the capital markets with an actual capital structure which is less than ideal. In addition, the hypothetical capital structure is based on a group of pipelines, thus representing the average level of risk in the industry. Consequently, the specific risks of TGPL relative to other pipelines are taken into account through adjusting the rate of return within the zone of reasonableness.

Given the complexity of the issue, FERC must flesh out its extremely terse analysis. FERC must more fully explain why its decision to allow a 14.45% rate of return does not amount to double dipping. In short, "the Commission 'crossed the line from the tolerably terse to the intolerably mute.' " TransCanada PipeLines Ltd. v. FERC, 24 F.3d 305, 310 (D.C.Cir.1994) (quoting Greater Boston Television Corp. v. FCC, 444 F.2d 841, 852 (D.C.Cir.1970), cert. denied, 403 U.S. 923, 91 S. Ct. 2233, 29 L. Ed. 2d 701 (1971)).

In addition, FERC's reasoning appears somewhat inconsistent with one of its earlier arguments. Here, FERC argues TGPL is placed at a disadvantage when the Commission imputes to it an ideal hypothetical capital structure because the pipeline itself must confront the markets with a less than ideal actual capital structure. Previously, however, FERC argued NCUC erred in asserting the "real" return on equity was 24% because NCUC relied on the discarded actual capital structure rather than the hypothetical structure which replaced it. We question how FERC can have it both ways. FERC must explain why it is not inconsistent for the hypothetical capital structure to be the measuring standard for one argument and not the other.

While accepting the Commission's decision to use a hypothetical capital structure, TGPL argues the Commission erred in relying on the primary comparison group of seven publicly listed parent companies with pipeline subsidiaries to develop the hypothetical capital structure. Rather, TGPL asserts the Commission should have imputed a capital structure for TGPL based on the average capital structure of the proxy group of similarly situated regulated pipelines. The only justification the Commission gave for its conclusion was that, unlike the regulated pipelines, the parent companies in the primary comparison group had publicly traded stocks. Thus, their capital structures were used to attract equity capital in the financial markets. 64 FERC at 61,345. The focus on publicly traded stock arises from the move to open-access transportation and the restructuring of the pipeline industry. TGPL alleges this rationale departs from the Commission's longstanding preference of relying on the actual capital structure of the pipeline to set its return rates, as long as the pipeline issues its own long-term debt to outside investors without any guarantees from its parent company. See Louisiana Intrastate Gas Co., 52 FERC p 61,297 at 62,188 (1990); Williams Pipeline Co., 31 FERC p 61,377 at 61,836 (1985); Midwestern Gas Transmission Co., 31 FERC p 61,317 at 61,720-21 (1985).

We agree that the Commission has not sufficiently explained its departure from its prior cases. While we concede the Commission's focus on publicly traded stock may be legitimate if the Commission were operating on a clean slate, this focus is manifestly inconsistent with the prior case law. The Commission futilely attempts to distinguish the case at bar on the basis that the case concerns the appropriate proxy group for developing a hypothetical capital structure once the Commission determines the parent's structure is inappropriate. In contrast, the cited cases address only the first step in the rate of return process: whether the capital structure of the pipeline or its parent should be used in determining a rate of return.

We find this a distinction without a difference. If TGPL had raised capital through the issuance of long-term debt, the Commission would have used TGPL's capital structure even though TGPL's stock is not publicly traded. FERC has not explained how the additional factor of a hypothetical capital structure changes the calculus, especially since FERC developed the fictional structure for TGPL, not its parent, TEC. In addition, for sixteen of the twenty pipelines in the secondary proxy group, the Commission uses their own capital structure in setting the rates of return in their respective rate cases. None of these sixteen issue their own stock. The Commission has not explained why it employs the capital structures of these non-publicly traded pipelines for setting their rates of return, but nonetheless finds these structures deficient for purposes of developing a hypothetical structure for TGPL. Consequently, the Commission has not provided a valid reason why this proxy group is not the most appropriate comparison group available in the record.2

Finally, TGPL argues even if the proxy group of parent pipelines is deemed valid, the Commission erred in including TEC in this group. We find this argument frivolous, warranting no discussion.

Because we remand to the Commission all but this last issue for fuller exposition, we have no need to address NCUC's denial of due process argument.

It is so ordered.